llvm-6502/test/CodeGen/ARM/2012-10-04-AAPCS-byval-align8.ll
Andrew Trick 6ea2b9608a Allocate local registers in order for optimal coloring.
Also avoid locals evicting locals just because they want a cheaper register.

Problem: MI Sched knows exactly how many registers we have and assumes
they can be colored. In cases where we have large blocks, usually from
unrolled loops, greedy coloring fails. This is a source of
"regressions" from the MI Scheduler on x86. I noticed this issue on
x86 where we have long chains of two-address defs in the same live
range. It's easy to see this in matrix multiplication benchmarks like
IRSmk and even the unit test misched-matmul.ll.

A fundamental difference between the LLVM register allocator and
conventional graph coloring is that in our model a live range can't
discover its neighbors, it can only verify its neighbors. That's why
we initially went for greedy coloring and added eviction to deal with
the hard cases. However, for singly defined and two-address live
ranges, we can optimally color without visiting neighbors simply by
processing the live ranges in instruction order.

Other beneficial side effects:

It is much easier to understand and debug regalloc for large blocks
when the live ranges are allocated in order. Yes, global allocation is
still very confusing, but it's nice to be able to comprehend what
happened locally.

Heuristics could be added to bias register assignment based on
instruction locality (think late register pairing, banks...).

Intuituvely this will make some test cases that are on the threshold
of register pressure more stable.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@187139 91177308-0d34-0410-b5e6-96231b3b80d8
2013-07-25 18:35:14 +00:00

64 lines
2.0 KiB
LLVM

; RUN: llc < %s -mtriple=armv7-none-linux-gnueabi | FileCheck %s
; Test that we correctly use registers and align elements when using va_arg
%struct_t = type { double, double, double }
@static_val = constant %struct_t { double 1.0, double 2.0, double 3.0 }
declare void @llvm.va_start(i8*) nounwind
declare void @llvm.va_end(i8*) nounwind
; CHECK-LABEL: test_byval_8_bytes_alignment:
define void @test_byval_8_bytes_alignment(i32 %i, ...) {
entry:
; CHECK: stm r0, {r1, r2, r3}
%g = alloca i8*
%g1 = bitcast i8** %g to i8*
call void @llvm.va_start(i8* %g1)
; CHECK: add [[REG:(r[0-9]+)|(lr)]], {{(r[0-9]+)|(lr)}}, #7
; CHECK: bfc [[REG]], #0, #3
%0 = va_arg i8** %g, double
call void @llvm.va_end(i8* %g1)
ret void
}
; CHECK-LABEL: main:
; CHECK: movw [[BASE:r[0-9]+]], :lower16:static_val
; CHECK: movt [[BASE]], :upper16:static_val
; ldm is not formed when the coalescer failed to coalesce everything.
; CHECK: ldrd r2, [[TMP:r[0-9]+]], {{\[}}[[BASE]]{{\]}}
; CHECK: movw r0, #555
define i32 @main() {
entry:
call void (i32, ...)* @test_byval_8_bytes_alignment(i32 555, %struct_t* byval @static_val)
ret i32 0
}
declare void @f(double);
; CHECK-LABEL: test_byval_8_bytes_alignment_fixed_arg:
; CHECK-NOT: str r1
; CHECK: str r3, [sp, #12]
; CHECK: str r2, [sp, #8]
; CHECK-NOT: str r1
define void @test_byval_8_bytes_alignment_fixed_arg(i32 %n1, %struct_t* byval %val) nounwind {
entry:
%a = getelementptr inbounds %struct_t* %val, i32 0, i32 0
%0 = load double* %a
call void (double)* @f(double %0)
ret void
}
; CHECK-LABEL: main_fixed_arg:
; CHECK: movw [[BASE:r[0-9]+]], :lower16:static_val
; CHECK: movt [[BASE]], :upper16:static_val
; ldm is not formed when the coalescer failed to coalesce everything.
; CHECK: ldrd r2, [[TMP:r[0-9]+]], {{\[}}[[BASE]]{{\]}}
; CHECK: movw r0, #555
define i32 @main_fixed_arg() {
entry:
call void (i32, %struct_t*)* @test_byval_8_bytes_alignment_fixed_arg(i32 555, %struct_t* byval @static_val)
ret i32 0
}